The aims of the Clinical Psychopharmacology Computer Laboratory are to provide quantitative and computer support for clinical psychopharmacology research through collaborative participation in projects conducted by investigators in our institutional setting and in other institutions as well. Different levels of involvement in different projects will extend from participation in actual data collection to the providing of specific specialized services such as computer classification of symptom profiles of patients who are research subjects. In addition to collaborative research support functions, activities supported by the grant will include (a) development and/or evaluation of measuring instruments and assessment devices having potential applications in clincal psychopharmacology research, (b) development and/or evaluation of diagnostic criteria and classification procedure for defining research populations, (c) development and/or evaluation of new statistical methods having potential applications in clinical psychopharmacology research. The methodology employed in pursuit of these aims will be statistical or psychometric. Standard statistical methods such as general linear model analysis of variance or covariance, as well as powerful quantitative models for ellucidation of multivariate relationships, are available for use in analysis of clinical data. Monte carlo methods will be relied on heavily in evaluating the performance of statistical tests under conditions simulating those encountered in clinical research, and exact methods will be used to calculate probabilities of Type I an Type II errors in other cases.